Predicting the impact of rare variants on RNA splicing in CAGI6.

Variants which disrupt splicing are a frequent cause of rare disease that have been under-ascertained clinically. Accurate and efficient methods to predict a variant's impact on splicing are needed to interpret the growing number of variants of unknown significance (VUS) identified by exome and genome sequencing. Here, we present the results of the CAGI6 Splicing VUS challenge, which invited predictions of the splicing impact of 56 variants ascertained clinically and functionally validated to determine splicing impact. The performance of 12 prediction methods, along with SpliceAI and CADD, was compared on the 56 functionally validated variants. The maximum accuracy achieved was 82% from two different approaches, one weighting SpliceAI scores by minor allele frequency, and one applying the recently published Splicing Prediction Pipeline (SPiP). SPiP performed optimally in terms of sensitivity, while an ensemble method combining multiple prediction tools and information from databases exceeded all others for specificity. Several challenge methods equalled or exceeded the performance of SpliceAI, with ultimate choice of prediction method likely to depend on experimental or clinical aims. One quarter of the variants were incorrectly predicted by at least 50% of the methods, highlighting the need for further improvements to splicing prediction methods for successful clinical application.

A novel variant in GATM causes idiopathic renal Fanconi syndrome and predicts progression to end-stage kidney disease.

Renal Fanconi syndrome (RFS) is a generalised disorder of the proximal convoluted tubule. Many genes have been associated with RFS including those that cause systemic disorders such as cystinosis, as well as isolated RFS. We discuss the case of a 10-year-old female who presented with leg pain and raised creatinine on a screening blood test. Her mother has RFS and required a kidney transplant in her thirties. Further investigations confirmed RFS in the daughter. Exome sequencing was performed on the affected mother, child, and unaffected father. We identified a novel variant in GATM; c.965G>C p.(Arg322Pro) segregating dominantly in the mother and daughter. We validated our finding with molecular dynamics simulations and demonstrated a dynamic signature that differentiates our variant and two previously identified pathogenic variants in GATM from wildtype. Genetic testing has uncovered a novel pathogenic variant that predicts progression to end stage kidney failure and has important implications for family planning and cascade testing. We recommend that GATM is screened for in children presenting with RFS, in addition to adults, particularly with kidney failure, who may have had previous negative gene testing.

Apparent Homozygosity for a gr/gr AZFc Deletion in A 47,XYY Man with Oligozoospermia and Secondary Infertility.

Background: Approximately 1 in 1000 men have a 47,XYY karyotype. Previous publications have presented cases of infertile XYY men and have suggested that the additional Y chromosome may cause disrupted meiosis leading to sperm apoptosis. The purpose of the current study was to determine whether XYY men are over-represented in infertility cohorts. Methods: In this paper, an ongoing infertility cohort was evaluated for Y chromosome microdeletions using the MLPA technique and the data from the first 2000 referrals were recorded. Moreover, the MLPA technique detected 47,XYY karyotypes. Results: Four XYY individuals were identified within the cohort. One of the four XYY men was shown to have an apparent gr/gr partial AZFc deletion on both Y chromosomes while Sertoli cell only syndrome was detected in another case. The other two cases (out of 2000) might, therefore, represent an incidental finding. Conclusion: The gr/gr deletion is not detectable by the multiplex PCR method; therefore, there might be additional explanations for the fertility problems of infertile XYY men reported in previously published articles. It seems that among other cases, their XYY karyotype may be coincidental, rather than causative of their fertility issues.

SHOX Whole Gene Duplications Are Overrepresented in SHOX Haploinsufficiency Phenotype Cohorts.

Transcription of SHOX is dependent upon the interaction of the gene with a complex array of flanking regulatory elements. Duplications that contain flanking regulatory elements but not the SHOX gene have been reported in individuals with SHOX haploinsufficiency syndromes, suggesting that alterations to the physical organisation or genomic architecture may affect SHOX transcription. Individuals with tall stature and an additional X or Y chromosome have an extra copy of both the SHOX gene and the entire SHOX regulatory region, so all three copies of SHOX can be expressed fully. However, for a duplication of the SHOX gene that does not include all of the flanking regulatory elements, the potential effect on SHOX expression is difficult to predict. We present nine unpublished individuals with a SHOX whole gene duplication in whom the duplication contains variable amounts of the SHOX regulatory region, and we review 29 similar cases from the literature where phenotypic data were clearly stated. While tall stature was present in a proportion of these cases, we present evidence that SHOX whole gene duplications can also result in a phenotype more typically associated with SHOX haploinsufficiency and are significantly overrepresented in Leri-Weill dyschondrosteosis and idiopathic short stature probands compared to population controls. Although similar-looking duplications do not always produce a consistent phenotype, there may be potential genotype-phenotype correlations regarding the duplication size, regulatory element content, and the breakpoint proximity to the SHOX gene. Although ClinGen does not currently consider SHOX whole gene duplications to be clinically significant, the ClinGen triplosensitivity score does not take into account the context of the duplication, and more is now known about SHOX duplications and the role of flanking elements in SHOX regulation. The evidence presented here suggests that these duplications should not be discounted without considering the extent of the duplication and the patient phenotype, and should be included in diagnostic laboratory reports as variants of uncertain significance. Given the uncertain pathogenicity of these duplications, any reports should encourage the exclusion of all other causes of short stature where possible.

AIFM1-associated X-linked spondylometaphyseal dysplasia with cerebral hypomyelination.

Spondylometaphyseal dysplasia with cerebral hypomyelination (SMD-H) is a very rare but distinctive phenotype, unusually combining spondylometaphyseal dysplasia with hypomyelinating leukodystrophy. Recently, SMD-H has been associated with variants confined to a specific intra-genic locus involving Exon 7, suggesting that AIFM1 plays an important role in bone development and metabolism as well as cerebral myelination. Here we describe two further affected boys, one with a novel intronic variant associated with skipping of Exon 7 of AIFM1 and the other a synonymous variant within Exon 7 of AIFM1. We describe their clinical course and radiological and genetic findings, providing further insight into the natural history of this condition.

Screening of a large Rubinstein-Taybi cohort identified many novel variants and emphasizes the importance of the CREBBP histone acetyltransferase domain.

Pathogenic variants within the CREBBP and EP300 genes account for the majority of individuals with Rubinstein-Taybi syndrome (RSTS). Data are presented from a large cohort of 395 individuals referred for diagnostic testing of CREBBP, and of the 19 CREBBP missense variants classified as likely pathogenic in this study, 17 were within the histone acetyltransferase (HAT) domain, providing evidence that this domain is critical to the normal function of the CREBBP protein (CBP). The data presented here, combined with other published results, suggest that the presence of a missense variant within the CBP HAT domain can be considered as moderate evidence of pathogenicity in the context of official variant interpretation guidelines. Within our study cohort, 129 had a pathogenic or likely pathogenic CREBBP variant and 5 had a variant of uncertain significance (VUS) which warranted familial studies. 147 of the remaining probands were also screened for EP300 and a further 16 pathogenic or likely pathogenic variants were identified, plus one VUS. Therefore, this analysis has provided a molecular diagnosis in at least 145 individuals with RSTS (37%) and identified a wide range of variants (n = 133) of which 103 were novel.

Screening of a large PAX6 cohort identified many novel variants and emphasises the importance of the paired and homeobox domains.

Pathogenic variants within PAX6 are most often associated with aniridia, but have been linked with other phenotypes such as nystagmus, cataracts and foveal hypoplasia. Data are presented from a large cohort of 434 probands referred for PAX6 diagnostic testing. This analysis identified a wide range of pathogenic variants (n = 145) in 254 probands (including 61 novel variants). Excluding missense variants predicted to affect splicing, all 29 of the remaining missense variants were located within the paired (n = 27) or homeobox (n = 2) domains of the PAX6 protein, providing further evidence that these domains are critical to normal PAX6 function. Genotype-phenotype evidence suggests that while aniridia is associated with most variant types, a much broader clinical spectrum is seen in patients harbouring a missense variant, or a frameshift or run-on variant that results in an elongated or extended PAX6 protein.

Directly Transmitted 12.3-Mb Deletion with a Consistent Phenotype in the Variable 11q21q22.3 Region.

A phenotype is emerging for the proximal pair of G-dark bands in 11q (11q14.1 and q14.3) but not yet for the distal pair (11q22.1 and q22.3). A mother and daughter with the same directly transmitted 12.3-Mb interstitial deletion of 11q21q22.3 (GRCh37: 93,551,765-105,817,723) both had initial feeding difficulties and failure to thrive, speech delay, learning difficulties, and mild dysmorphism. Among 17 patients with overlapping deletions, developmental or speech delay, dysmorphism, hypotonia, intellectual disability or learning difficulties, short stature, and coloboma were each found in 2 or more. These results may provide the basis for a consistent phenotype for this region. Among the 53 deleted and additional breakpoint genes, CNTN5, YAP1, and GRI4 were the most likely candidates. Non-penetrance of haploinsufficient genes and dosage compensation among related genes may account for the normal cognition in the mother and variable phenotypes that can extend into the normal range.

Correction: Blood RNA analysis can increase clinical diagnostic rate and resolve variants of uncertain significance.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Blood RNA analysis can increase clinical diagnostic rate and resolve variants of uncertain significance.

PURPOSE: Diagnosis of genetic disorders is hampered by large numbers of variants of uncertain significance (VUSs) identified through next-generation sequencing. Many such variants may disrupt normal RNA splicing. We examined effects on splicing of a large cohort of clinically identified variants and compared performance of bioinformatic splicing prediction tools commonly used in diagnostic laboratories. METHODS: Two hundred fifty-seven variants (coding and noncoding) were referred for analysis across three laboratories. Blood RNA samples underwent targeted reverse transcription polymerase chain reaction (RT-PCR) analysis with Sanger sequencing of PCR products and agarose gel electrophoresis. Seventeen samples also underwent transcriptome-wide RNA sequencing with targeted splicing analysis based on Sashimi plot visualization. Bioinformatic splicing predictions were obtained using Alamut, HSF 3.1, and SpliceAI software. RESULTS: Eighty-five variants (33%) were associated with abnormal splicing. The most frequent abnormality was upstream exon skipping (39/85 variants), which was most often associated with splice donor region variants. SpliceAI had greatest accuracy in predicting splicing abnormalities (0.91) and outperformed other tools in sensitivity and specificity. CONCLUSION: Splicing analysis of blood RNA identifies diagnostically important splicing abnormalities and clarifies functional effects of a significant proportion of VUSs. Bioinformatic predictions are improving but still make significant errors. RNA analysis should therefore be routinely considered in genetic disease diagnostics.

Screening of a large cohort of blepharophimosis, ptosis, and epicanthus inversus syndrome patients reveals a very strong paternal inheritance bias and a wide spectrum of novel FOXL2 mutations.

Blepharophimosis, Ptosis, and Epicanthus inversus Syndrome (BPES) is caused by autosomal dominant mutations in FOXL2. There are two forms of BPES: type I (with primary ovarian insufficiency (POI)) and type II (without POI). Data are presented from a large cohort of 177 BPES probands. Diagnostic testing identified a wide range of mutations in 119 mutation-positive patients (including 38 novel mutations). Although FOXL2 mutations are distributed throughout the gene, over 50% were frameshift mutations within a hotspot region of the gene that can be detected using a single primer pair to provide a cost-effective and rapid screening method. There was a significant proportion of de novo cases in this study, although in 7% there may be undetected parental mosaicism. There was an excess of female compared to male probands and a highly significant bias in the parental original of inherited mutations, with 20/21 found to be paternal in origin (95%). This could be because BPES in a female is more likely to come to clinical attention and because there is a generalised and more widespread clinical effect on fertility, in addition to the established association with POI. This study demonstrates the importance of cascade screening and provides new information on inheritance and parental mosaicism in BPES which will aid genetic counselling and accurate risk management.

Autosomal dominant tubulointerstitial kidney disease-UMOD is the most frequent non polycystic genetic kidney disease.

BACKGROUND: Autosomal dominant tubulointerstitial kidney disease (ADTKD) caused by mutations in the UMOD gene (ADTKD-UMOD) is considered rare and often remains unrecognised. We aimed to establish the prevalence of genetic kidney diseases, ADTKD and ADTKD-UMOD in adult chronic kidney disease (CKD) patients, and to investigate characteristic features. METHODS: We sent questionnaires on family history to all patients with CKD stages 3-5 in our tertiary renal centre to identify patients with inherited renal disease. Details on clinical and family history were obtained from patient interviews and clinical records. Sanger sequencing of the UMOD gene was performed from blood or saliva samples. RESULTS: 2027 of 3770 sent questionnaires were returned. 459 patients reported a family history, which was consistent with inherited kidney disease in 217 patients. 182 non-responders with inherited kidney diseases were identified through a database search. Of these 399 individuals, 252 had autosomal dominant polycystic kidney disease (ADPKD), 28 had ADTKD, 25 had Alports, and 44 were unknown, resulting in 11% of CKD 3-5 patients and 19% of end-stage renal disease patients with genetic kidney diseases. Of the unknown, 40 were genotyped, of whom 31 had findings consistent with ADTKD. 30% of unknowns and 39% of unknowns with ADTKD had UMOD mutations. Altogether, 35 individuals from 18 families were found to have ten distinct UMOD mutations (three novel), making up 1% of patients with CKD 3-5, 2% of patients with end-stage renal disease, 9% of inherited kidney diseases and 56% with ADTKD. ADTKD-UMOD was the most common genetic kidney disease after ADPKD with a population prevalence of 9 per million. Less proteinuria and haematuria, but not hyperuricaemia or gout were predictive of ADTKD-UMOD. The main limitations of the study are the single-centre design and a predominantly Caucasian population. CONCLUSIONS: The prevalence of genetic kidney diseases and ADTKD-UMOD is significantly higher than previously described. Clinical features poorly predicted ADTKD-UMOD, highlighting the need for genetic testing guided by family history alone.

Incomplete penetrance, variable expressivity, or dosage insensitivity in four families with directly transmitted unbalanced chromosome abnormalities.

The direct transmission of microscopically visible unbalanced chromosome abnormalities (UBCAs) is rare and usually has phenotypic consequences. Here we report four families in which a normal phenotype was initially found in one or more family members. Each UBCA was interpreted with regard to overlapping examples and factors previously associated with transmitted imbalances including incidental ascertainment, low gene density, benign copy number variation (CNV) content, and gene relatedness. A 4.56 Mb deletion of 8p23.1-p23.2 was thought to be causal in the affected proband but showed incomplete penetrance in her mother and sibling (Family 1). Incomplete penetrance was also associated with a 10.88 Mb duplication of 13q21.31-q22.1 (Family 3) and dosage insensitivity with a 17.6 Mb deletion of 22pter-q11.21 (Family 4) that were both ascertained at prenatal diagnosis and each found in 4 unaffected family members. The 22pter-q11.21 deletion is part of a region with high benign CNV content and supports the mapping of cat eye syndrome to a 600 kb interval of 22q11.1-q11.21. Low gene densities of less than 2.0 genes/Mb were found in each of these three families but only after segmentally duplicated genes were excluded from the deletions of 8p and 22q. In contrast, gene density was average and variable expressivity associated with a 3.59 Mb duplication of 8p23.1 incidentally ascertained for paternal infertility (Family 2). Our results indicate that a greater degree of direct parental transmission, incomplete penetrance, and variable expression are features of both sub-microscopic CNVs and UBCAs with relatively low gene and high benign CNV content.

Collagen (COL4A) mutations are the most frequent mutations underlying adult focal segmental glomerulosclerosis.

BACKGROUND: Multiple genes underlying focal segmental glomerulosclerosis (FSGS) and/or steroid-resistant nephrotic syndrome (SRNS) have been identified, with the recent inclusion of collagen IV mutations responsible for Alport disease (AD) or thin basement membrane nephropathy (TBMN). We aimed to investigate the distribution of gene mutations in adult patients with primary FSGS/SRNS by targeted next generation sequencing (NGS). METHODS: Eighty-one adults from 76 families were recruited; 24 families had a history of renal disease. A targeted NGS panel was designed and applied, covering 39 genes implicated in FSGS/SRNS including COL4A3-5. RESULTS: Confirmed pathogenic mutations were found in 10 patients (6 with family history) from 9 families (diagnostic rate 12%). Probably pathogenic mutations were identified in an additional six patients (combined diagnostic rate 20%). Definitely pathogenic mutations were identified in 22% of patients with family history and 10% without. Mutations in COL4A3-5 were present in eight patients from six families, representing 56% of definitely pathogenic mutations, and establishing a diagnosis of AD in six patients and TBMN in two patients. Collagen mutations were identified in 38% of families with familial FSGS, and 3% with sporadic FSGS, with over half the mutations occurring in COL4A5. Patients with collagen mutations were younger at presentation and more likely to have family history, haematuria and glomerular basement membrane abnormalities. CONCLUSIONS: We show that collagen IV mutations, including COL4A5, frequently underlie FSGS and should be considered, particularly with a positive family history. Targeted NGS improves diagnostic efficiency by investigating many candidate genes in parallel.

Duplications upstream and downstream of SHOX identified as novel causes of Leri-Weill dyschondrosteosis or idiopathic short stature.

Leri-Weill dyschondrosteosis is a pseudoautosomal dominantly-inherited skeletal dysplasia ascribed to haploinsufficiency of the SHOX gene caused by deletions, point mutations, or partial duplications of the gene, or to heterozygous deletions upstream or downstream of the intact SHOX gene involving conserved non-coding cis-regulatory DNA elements that show enhancer activity. Recently, two SHOX conserved non-coding element duplications, one upstream and one downstream, were reported in patients referred with idiopathic short stature. To further evaluate the role of these duplications in SHOX-related disorders, we describe seven patients (five with Leri-Weill dyschondrosteosis and two with short stature) all of whom have duplications of part of the upstream or downstream conserved non-coding element regions, identified by multiplex ligation-dependent probe amplification. In addition, we show data from 32 patients with an apparently identical downstream duplication that includes a proposed putative regulatory element (identified by multiplex ligation-dependent probe amplification or array comparative genome hybridization), which results in a variable phenotype from normal to mild Leri-Weill dyschondrosteosis. These additional data provide further evidence that duplications of upstream and downstream long range cis-regulatory DNA elements can result in a SHOX-related phenotype.

Homozygosity for a novel deletion downstream of the SHOX gene provides evidence for an additional long range regulatory region with a mild phenotypic effect.

Léri-Weill dyschondrosteosis is caused by heterozygous mutations in SHOX or its flanking sequences, including whole or partial gene deletions, point mutations within the coding sequence, and deletions of downstream regulatory elements. The same mutations when biallelic cause the more severe Langer Mesomelic dysplasia. Here, we report on a consanguineous family with a novel deletion downstream of SHOX in which homozygously deleted individuals have a phenotype intermediate between Léri-Weill dyschondrosteosis and Langer Mesomelic dysplasia while heterozygously deleted individuals are mostly asymptomatic. The deleted region is distal to all previously described 3' deletions, suggesting the presence of an additional regulatory element, deletions of which have a milder, variable phenotypic effect.

Diagnostic screening identifies a wide range of mutations involving the SHOX gene, including a common 47.5 kb deletion 160 kb downstream with a variable phenotypic effect.

Léri-Weill dyschondrosteosis (LWD) results from heterozygous mutations of the SHOX gene, with homozygosity or compound heterozygosity resulting in the more severe form, Langer mesomelic dysplasia (LMD). These mutations typically take the form of whole or partial gene deletions, point mutations within the coding sequence, or large (>100 kb) 3' deletions of downstream regulatory elements. We have analyzed the coding sequence of the SHOX gene and its downstream regulatory regions in a cohort of 377 individuals referred with symptoms of LWD, LMD or short stature. A causative mutation was identified in 68% of the probands with LWD or LMD (91/134). In addition, a 47.5 kb deletion was found 160 kb downstream of the SHOX gene in 17 of the 377 patients (12% of the LWD referrals, 4.5% of all referrals). In 14 of these 17 patients, this was the only potentially causative abnormality detected (13 had symptoms consistent with LWD and one had short stature only), but the other three 47.5 kb deletions were found in patients with an additional causative SHOX mutation (with symptoms of LWD rather than LMD). Parental samples were available on 14/17 of these families, and analysis of these showed a more variable phenotype ranging from apparently unaffected to LWD. Breakpoint sequence analysis has shown that the 47.5 kb deletion is identical in all 17 patients, most likely due to an ancient founder mutation rather than recurrence. This deletion was not seen in 471 normal controls (P<0.0001), providing further evidence for a phenotypic effect, albeit one with variable penetration.

Exome analysis resolves differential diagnosis of familial kidney disease and uncovers a potential confounding variant.

A girl aged 6 presented with haematuria and her sister (aged 5) presented with haematuria and proteinuria. Family history showed multiple individuals suffering from end stage renal failure from the paternal side of the pedigree. Following kidney biopsy in the father and paternal grandmother, the pathological diagnosis was of focal segmental glomerulosclerosis (FSGS). Exome sequencing was undertaken in the proband's sister and grandmother. Genetic variants shared by both affected individuals were interrogated to identify the genetic cause of disease. Candidate variants were then sequenced in all the family members to determine segregation with the disease. A mutation of COL4A5 known to cause Alport syndrome segregated with disease from the paternal side of the pedigree and a variant in NPHS1 was present in both paediatric cases and inherited from their mother. This study highlights the advantages of exome sequencing over single gene testing; disease presentation can be heterogeneous with several genes representing plausible candidates; candidate gene(s) may be unavailable as a diagnostic test; consecutive, single gene testing typically concludes once a single causal mutation is identified. In this family, we were able to confirm a diagnosis of Alport syndrome, which will facilitate testing in other family members.

Detection of Partial Deletions of Y-chromosome AZFc in Infertile Men Using the Multiplex Ligation-dependent Probe Amplification Assay.

BACKGROUND: In recent studies, partial deletions of the azoospermia factor c region (AZFc) on the Y-chromosome have been detected in males with infertility problems. However, there has been a lot of debate about their significance. In order to study such deletions, a simple but accurate method for their detection was applied in this study. METHODS: We present data obtained from the Multiplex Ligation-dependent Probe Amplification (MLPA) assay using a new Y-chromosome-specific MLPA probemix (P360) which allows the easy detection of partial AZFc deletions. RESULTS: Partial AZFc deletions were detected in 8% of our cohort of previously mutation-negative infertile males (and 0% of the fertile control cohort). CONCLUSION: These results provide further evidence of the causality of partial AZFc deletions. None of the partial AZFc deletions were detectable by the standard multiplex PCR method, demonstrating the advantage of the MLPA method.

Recurrence and variability of germline EPCAM deletions in Lynch syndrome.

Recently, we identified 3' end deletions in the EPCAM gene as a novel cause of Lynch syndrome. These truncating EPCAM deletions cause allele-specific epigenetic silencing of the neighboring DNA mismatch repair gene MSH2 in tissues expressing EPCAM. Here we screened a cohort of unexplained Lynch-like families for the presence of EPCAM deletions. We identified 27 novel independent MSH2-deficient families from multiple geographical origins with varying deletions all encompassing the 3' end of EPCAM, but leaving the MSH2 gene intact. Within The Netherlands and Germany, EPCAM deletions appeared to represent at least 2.8% and 1.1% of the confirmed Lynch syndrome families, respectively. MSH2 promoter methylation was observed in epithelial tissues of all deletion carriers tested, thus confirming silencing of MSH2 as the causative defect. In a total of 45 families, 19 different deletions were found, all including the last two exons and the transcription termination signal of EPCAM. All deletions appeared to originate from Alu-repeat mediated recombination events. In 17 cases regions of microhomology around the breakpoints were found, suggesting nonallelic homologous recombination as the most likely mechanism. We conclude that 3' end EPCAM deletions are a recurrent cause of Lynch syndrome, which should be implemented in routine Lynch syndrome diagnostics.